Publications | Department of Knowledge Technologies https://kt.ijs.si Tue, 08 Sep 2026 12:12:33 +0000 en-US hourly 1 https://wordpress.org/?v=7.1 https://kt.ijs.si/wp-content/uploads/2021/04/cropped-Knowledge-tehnologies_favicon_s-32x32.png Publications | Department of Knowledge Technologies https://kt.ijs.si 32 32 Rapid evidence mapping of soil fauna responses to agricultural management assisted by large language models https://kt.ijs.si/news/rapid-evidence-mapping-of-soil-fauna-responses-to-agricultural-management-assisted-by-large-language-models/ Sat, 08 Aug 2026 12:10:16 +0000 https://kt.ijs.si/?p=7506 A new study, Rapid evidence mapping of soil fauna responses to agricultural management assisted by large language models, published in Geoderma examines how artificial intelligence can assist researchers in navigating the growing volume of scientific literature.

The authors, including Marko Debeljak and Vid Podpečan from the EcoEnvAI research group, from E8 at the Jožef Stefan Institute, alongside an international team of collaborators, developed an LLM-assisted workflow to automatically extract and structure knowledge from scientific abstracts, enabling rapid evidence mapping in soil science.

The study demonstrates the application of AI in supporting environmental research, identifying knowledge gaps, and accelerating scientific synthesis.

The open-access paper is available at: https://www.sciencedirect.com/science/article/pii/S0016706126002181

]]> FuDoBa has been published in the Machine Learning Journal by Springer https://kt.ijs.si/news/fudoba-has-been-published-in-the-machine-learning-journal-by-springer/ Mon, 03 Aug 2026 09:41:50 +0000 https://kt.ijs.si/?p=7490 📄 “FuDoBa: Fusing Document and Knowledge Graph Based Representations with Bayesian Optimisation” with Boshko Koloski, Senja Pollak, Roberto Navigli and Blaž Škrlj, has been published by Springer,

This framework improves off-the-shelf LLM embeddings (such as OpenAI text-embedding) for classification tasks by incorporating global knowledge from Wikidata KGs and local knowledge learned from task-specific knowledge graphs. High-dimensional fusion can be matched by searching for low-dimensional weighted modality contributions via Bayesian Optimization. Furthermore, tabular foundation models serve as effective surrogates, achieving results comparable to more expensive AutoML runs at a fraction of the computational cost.

📄 Link: https://lnkd.in/dUmENV-e

]]>